用 Codex 或 Claude 帮你安装 复制这段 Prompt,粘贴到 Codex、Claude 或其他助手里,让它检查 Skill 页面并帮你完成安装。
直接命令不会经过审查 Prompt;运行前请先检查来源。
npx skills add https://github.com/benjam3n/reasoningtool --skill obo-obvious-bad-outcomes命令会保持在同一行。复制前请横向滚动并检查完整内容。
想先保存到本地?可下载 SkillsMP 当前能够提供的文件。
基于 SOC 职业分类
正在显示 SKILL.md
| name | obo - Obvious Bad Outcomes |
| description | Check for obvious bad outcomes that are being ignored, minimized, or rationalized away. |
| output | {"format":"prose"} |
Input: $ARGUMENTS
The most dangerous risks aren't the subtle ones — they're the obvious ones that everyone sees but nobody says out loud. Optimism bias, groupthink, and sunk costs cause people to rationalize away clear dangers. This skill forces the uncomfortable obvious onto the table.
What is being analyzed?
SUBJECT: [one sentence]
For each category, identify what bad thing is obviously true or obviously likely:
Ask directly:
If this goes wrong in the most obvious way, and someone asks "didn't anyone see this coming?" — what would the answer be?
If the answer is "yes, everyone saw it coming" — that's the finding.
OBVIOUS BAD OUTCOMES:
Elephant in the room:
- [the thing everyone knows but isn't saying]
Most likely failure:
- [the obvious way this goes wrong]
Predictable pattern:
- [how this type of thing usually fails]
Who gets hurt:
- [who bears the downside]
Most dangerous overlooked risk: [the one being actively rationalized away]
What honest acknowledgment changes: [how the analysis shifts when you stop ignoring this]
Use with:
/ogo -> Balance with obvious good outcomes/saf -> Full safety analysis/obv -> General obvious check/oba -> Obvious everything scan/prm -> Pre-mortem for detailed failure analysis